HD 189733 and the area around, with the nebula M27 on the right. The reference stars used are marked "R".
Not many observations have been reported on the internet after the first detection of the passage, and I believe that this is the reason why the transit window stated for the 10th of August 2006 at www.transitsearch.org had an uncertainty of almost 6 hours, lasting from 04.21 until 10.04 UTC. During this period I did measurements about three times per hour with the "Areo4" telescope and a photometric V-band filter. A few days later I found a more recent observation from July 30 this year on the web. By comparing the Julian dates of that observation with the date of the first discovery I calculated that the passage should have lasted from 05.07 until 06.42 UTC (±10 minutes) at the date of my observation. This is in good agreement with my observations below, where one clearly can see the start and the end of the passage. Unfortunately, most of the measurements made during the actual passage had to be rejected due to overexposure. I had forgotten that the star would get brighter as it climbed higher above the horizon after the measurements were started.
Changes in brightness from the star HD 189733 while a planet on the size of Jupiter passes in front. Unfortunately only successful measurements were only collected near the start and the end of the passage.
Each point in the figure represents the mean intensities extracted from three exposures lasting 12-16 seconds each. The telescope was moved slightly between each exposure to average out possible variations in the CCD-response (or more precisely, errors in the flat exposure). The intensity of HD189733 is calculated relative to four reference stars using the photometry tool in MaximDL. The script that controlled the exposures needed about 2.5 minutes to position the telescope and take the three exposures used to generate one point in the plot.
The standard deviation of the measurements before and after the transit is only 0.3 % (0.003 magnitudes), which is very good! The number of detected photons per point in the graph is roughly n1 = 106 from HD189733 and 1.5·106 combined from the reference stars. The quantum noise should then contribute with a standard deviation of (1/n1+1/n2)1/2 = 0.13 % in the measured values. Consequently, there must be other error sources involved as well. A probable candidate is scintillation noise, which is the same effect that makes the stars twinkle when we look at them.
I plan to do new measurements on this star to get more measurement points during the actual passage. It is also possible to do measurements on the ”candidate stars” at www.transitsearch.org . These are stars where one suspect that it is possible to detect a planet passage with photometry.
erlend.web&eronn.net [replace & with @]